Saturday, July 18, 2015

Utah and Alaska are on course to expand Medicaid. Are their eligibility estimates inflated? (Updated)

Both Alaska and Utah seem to be on course to implement the ACA Medicaid expansion. In Alaska, Governor Bill Walker is asserting his authority to do so without the approval of the legislature. In Utah, Governor Gary Herbert has struck a deal with legislative leaders -- still perhaps facing a rocky course toward legislative approval -- to seek a waiver for a "private option" form of the expansion.

In both states, as in any state, pro and con arguments over expansion are based on cost estimates which are in turn based on estimates of how many state residents will gain eligibility for Medicaid. It's noteworthy that in both states, in 2015, the Kaiser Family Foundation radically cut its 2014 estimate of those in the "coverage gap" -- that is, those whose incomes would have qualified them for Medicaid had their states accepted the expansion but are too low to qualify them for subsidized private plan coverage on ACA exchanges.*

In Alaska, Kaiser's estimate of the Medicaid gap population shrank from 17,290 as of March 31, 2014 to 10,500 as of  April 17 of this year. In Utah, the drop was from 57,850 in 2014 to 30,000 in 2015. These estimate cuts are mainly due to a redesign of insurance questions in the Census Bureau's Current Population Survey, intended to address previous under-reporting of respondents' insurance coverage.** While Kaiser cut its estimates for every state, the reductions in Utah and Alaska were far and away the largest. In both states, though, state officials' estimates of the "gap" population are closer to Kaiser's earlier estimates than to the newly reduced ones. In both states, in fact, working estimates are higher than Kaiser's 2014 figures.

In Utah, opponents of Governor Herbert's Healthy Utah plan, which would use federal Medicaid expansion funds to fund private plan coverage for the Medicaid-eligible, advanced a much more limited state-funded plan that would cover primary care visits but not specialty care or hospitalization. Supporters of the governor's plan commissioned a study by Notalys which found that the Medicaid expansion would be six times as cost-effective for the state. That study cites a previous Notalys consulting report from 2014 estimating the coverage gap population at 66,000, I can't locate that document, but in July 2014 Notalys released a widely covered study that pegged the number of those in the coverage gap (under 100% FPL and uninsured) at 77,127. That study claims that its methodology "was closely patterned after that used by the Kaiser Family Foundation (KFF) in its estimates of the coverage gap using the same CPS surveys." True enough, but that was in 2014, and like Kaiser's 2014 analysis, it's based on the 2012 and 2013 CPS surveys. Prior to that study's release, Governor Herbert relied on an early study that pegged the coverage gap at 56,000 -  close to Kaiser's 2014 estimate. The currently posted plan, released in Dec. 2014, anticipates covering 63,000 adults under 100% FPL.

The Healthy Alaska plan published by Alaska's Dept. of Health and Human Services relies for its Medicaid expansion population estimates primarily on an analysis by Evergreen Economics, published in February 2015. Evergreen estimates the coverage gap population -- adults aged 19-64 newly eligible for Medicaid and under 100% FPL -- at 23,344, more than double Kaiser's current estimate of 10,500. Evergreen did not rely on the federal CPS but rather on the state Public Health Division's Behavioral Risk Factor Surveillance System (BRFSS) survey for 2012 and 2013 and projections reported by the state's department of labor.*** The BRFSS surveys in both years asked respondents whether they were insured; the 2013 survey asked what type of insurance they had. Whether Evergreen broke the population into taxable household units comparable to Kaiser's, and designed to match eligibility criteria for Medicaid, I don't know.

In Utah, where Kaiser dropped its estimate of the Medicaid gap population  by 48%, from 57,850 in 2014 to 30,000 this year, I'd like to stick an oar in with a very crude measure of my own. In every state that accepts the Medicaid expansion belatedly, a portion of the newly Medicaid-eligible population, those with incomes ranging from 100% to 138% FPL, has to that point been eligible for subsidized private coverage purchased in the ACA exchanges. Unfortunately we don't know exactly how many ACA private plan customers fall into this category. But  HHS does report how many exchange customers have incomes in a somewhat wider range, 100-150% FPL. In nonexpansion states, as of mid-February, 2.85% million**** were in that category (probably two thirds to three quarters of them were in the 100-138% shoulda-been-in-Medicaid range, but that's immaterial to this guesstimate.) Kaiser, meanwhile, estimates 3.71 million in the coverage gap (under 100% FPL and ineligible for Medicaid in nonexpansion states). The estimated coverage gap population, then, is 1.3 times the private plan buyership in the 100-150% FPL range in the gap states. Now, Utah had 44,155 buyers between 100% and 150% FPL as of mid-February. Multiply by 1.3, and you get about 58,800 in the Medicaid gap -- very close to Kaiser's 2014 estimate [BUT SEE UPDATE/CORRECTION AT BOTTOM].

This may be a somewhat silly exercise, as the working assumption is that Kaiser's 50-state estimate for the Medicaid gap is accurate but that something might be funky in its Utah estimate. Also, Utah is wealthy state, so the ratio of its private plan customer base in the 100-150% FPL band to its population under 100% FPL may be higher than in most nonexpansion states, most of which are low-income states. (The state's percentage of target population enrolled in private plans, as estimated by Kaiser, is close to the national average, so that factor shouldn't skew this estimate.) If I can get some insight as to why Kaiser's gap estimates for Utah and Alaska dropped so precipitously, I will report.

UPDATE, 7/19: It just occurred to me that the percentage of Utah private plan buyers in the 100-150% FPL band who are actually between 100% and 138% FPL (and so potentially eligible for Medicaid)  is much lower than I was thinking. That's because the overall percentage of Utah buyers in the 100-150% FPL range looks more like the percentage in expansion than in nonexpansion states -- probably because Utah is a relatively wealthy state, unlike most nonexpansion states. In Utah, 31% of all buyers were 100-150% FPL. In nonexpansion states, 47% of buyers were in that range; in expansion states, just 22% were -- and all of them were actually 138-150% FPL, since below 138% they are eligible for Medicaid and ineligible for private plan subsidies. So let's assume for the moment that if there were no buyers under 138% FPL in Utah, those in the 138-150% band would be 22% of the whole. That would mean 27,205 out 123,662. Assuming then (a large assumption, I know) that there were 27,205 buyers between 138% and 150% FPL as of mid-February, there were also 16,950 shoulda-been-Medicaid-eligible buyers in the 100-138% FPL range. That's where my error may have been: I was assuming about twice as many 100-138% FPL-ers, because in a previous calculation similar to this one I determined that about two thirds of buyers in the 100-150% FPL range in nonexpansion states were in the 100-138% FPL range.

By that measure, about 1.9 million private plan enrollees in nonexpansion states should have been Medicaid eligible (i.e. were in the 100-138% FPL income range). That's slightly more than half the number of people that Kaiser estimates to be in the Medicaid gap. If that ratio holds for Utah, and my estimate of roughly 17k 100-138% FPLers is close to the mark (two big ifs), that suggests about 34,000 in the Medicaid gap -- close enough for government work to Kaiser's reduced estimate of 30,000.

One last guesstimate: Avalere Health calculated that 76% of uninsured Americans in the 100-150% FPL range who were eligible for subsidized private plans did in fact buy them. While Avalon could not cut into the 100-150% FPL range to determine the takeup rate among those within incomes from 100-138% FPL, it was almost certainly higher than the rate for the broader band, since premiums for  the benchmark silver plan for those under 138% FPL are capped at 2% of income -- and generally, the higher the income, the lower the takeup rate. If, the roughly 17,000 Utah buyers who should become Medicaid-eligible were, say, 80% of uninsured Utahans in that 100-138% FPL income range, there are perhaps 21,000-plus Medicaid-eligibles in that upper income range. Added to 34,000 below 100% FPL, that suggests about 55,000 people newly eligible for Medicaid in Utah.

UPDATE 2, 7/20/15: The method for estimating Medicaid gappers laid in update 1 above does not work for Alaska or Virginia -- two more wealthy states where the percentage of QHP buyers in the 100-150% FPL band looks more like the average for expansion than for nonexpansion states. There are just too many variables for a ratio like that to mean much, I guess -- including wide variations in the income distributions of QHP buyers in each state, the percentage of subsidy-eligible uninsured residents who bought QHPs, etc.   So I think the only real takeaways here are as follows: 1) government officials and commissioned researchers in Utah and Alaska are anticipating a far greater expansion population than Kaiser is -- effectively discounting the more rigorous questioning about insurance status in the latest CPS; 2) the income distribution of QHP buyers in wealthy nonexpansion states looks more like the distribution in expansion states (which are generally wealthier) than in the average nonexpansion state; 3) in small states like Alaska and Utah, eligibility estimates are very uncertain no matter who's doing them; and 4) HHS could clarify an awful lot if they would provide stats as to how many QHP buyers in nonexpansion were in the 100-138% FPL (Medicaid eligible) income range.

----

* The coverage gap refers to those whose household incomes are below 100% of the Federal Poverty Level (FPL) in states that refused the expansion and who are ineligible for Medicaid according to the state's existing criteria. In expansion states, those with incomes up to 138% FPL are eligible for Medicaid; in nonexpansion states, eligibility for subsidized private plans begins at 100% FPL.

** Here is Kaiser's explanation of the reduced estimates, from the 2015 coverage gap report linked to above:
This analysis uses data from the 2014 Current Population Survey (CPS) Annual Social and Economic Supplement (ASEC). The CPS ASEC provides socioeconomic and demographic information for the United Sates population and specific subpopulations. Importantly, the CPS ASEC provides detailed data on families and households, which we use to determine income for ACA eligibility purposes (see below for more detail).  Notably, with the 2014 ASEC, Census implemented a fundamental redesign of the health insurance coverage questions. This redesign aimed both to address longstanding issues with measurement of insurance coverage in the ASEC and to capture new coverage categories available under the ACA. The redesigned insurance questions lead to a lower estimate of the uninsured rate compared to the previous approach, addressing a longstanding issue of under-reporting of coverage in the ASEC.  As a result of these changes, health coverage data for the 2014 release (reflecting coverage in calendar year 2013) are not comparable with estimates from previous years.
*** The Alaska Department of Labor and Workforce Development (ADLWD).

**** HHS only breaks out income for states using hc.gov, which all nonexpansion states with the exception of Idaho do. I've added 40,000 private plan buyers at 100-150% FPL for Idaho -- a bit less than half its total private plan buyership, in line with other nonexpansion states.

No comments:

Post a Comment